new dcm for spectroscopy an engineering...new dcm for spectroscopy an engineering challenge -13 may...

$: New DCM for spectroscopy an engineering...New DCM for spectroscopy an engineering challenge -13 May 2014 - Y. Dabin The scattered power (fraction of energy irradiated) has been calculated$

New DCM for spectroscopy an engineering challenge review

Y. Dabin , R. Baker, L. Zhang, H. Gonzalez, R. Barrett, Ph. Marion, O. Mathon, R. Tucoulou

New DCM for spectroscopy an engineering challenge - 13 May 2014 - Y. DabinPage 2

DRIVERS FOR A NEW DCM: GENERAL OUTLINES


THE CRYSTALS CONFIGURATIONS

1st /2ndXtalOn axis

Crystals stroke

Axisoffset


CRYSTAL POSITIONING: PLACING THE CENTRE OF ROTATION

1st Xtal “fixed”

d

CoR White beam spot, moving or fixed on theFirst crystal:Fixed: thermal time constant better “averaged”

T

white

d

1st Xtal “moving”

BUT:

Bragg radial stroke

Bragg longitudinal stroke

Independent of configuration

B

dT

θcos2=

B

dL

θsin2=

T

L

T

L


d

CoR

Monochromatic beam “fixed” on 2nd Xtal:2nd Xtal Bragg angle better defined

1st crystal moving: High flow LN2 set

Twhite

CRYSTAL POSITIONING – CENTRE OF ROTATION LOCATION

)tan(1

minx

B L

hba

+=θ

Centre of rotation

white

Placing the white beam out of the axisTricky thermalConditionsAt max Bragg

white


b

a

h

At max Bragg

white

A LOT OF CRYSTALS COMBINATIONS POSSIBLE

Better thermalconditions

�White beam at CoR�CoR on the crystal surface

� Hot spot constant location

conditions

Possibility of making hot spot movingReducing the minimum Bragg

Parea=maximum

B

dT

θcos2=

unchanged

See L. Zhang’s talk


Possibility of making hot spot movingReducing the minimum Bragg

B

dL

θsin2=

~ Xtal length

unchanged

THE CRYSTAL PARALLELISM : THE FUNDAMENTAL ISSUE

Cryomount

AngularCorrectioncontrol

Gonios&tables


Crystal parallelism is the fundamental performance of any DCM

tables

CRYSTALS PARALLELISM DRIVERS

Drivers items characteristic cure Studies drivers

GeometricMoving stages ~ repeat Metrology-

lookup tablePiezos

Static studiesPlate deformation ~ repeat Piezosdeformation ~ repeat

Cryo-field from crystal set Stage shrinkage instability Thermal break

heater (TTF) Thermal studies&SimulatingThermal transferFunction

Thermal field from scattering

Stages and holders instability Radiation shield

Heater (TTF)

Crystals at different temp.

Crystal Cryo-box

Lack of cooling power rate.Response time

LN2 on 2nd

crystal

Dynamics Moving stagesFlexures Fast events

Push dynamicstudies-high frequency track

Vibration analysis


frequency track

CRYSTAL CAGE PARALLELISM : THE ROTATING PLATE AND TRAVEL CASES

Angular deviationFrom a lack of Stiffness(Bragg dependent)

Rotating plateDrives angular errors

Depends onBall track straightness

Best figures are: 25 µrd/200mm

Probably then : 0.1 µrd/mm (locally?)

(Bragg dependent)

20 Kg


Probably then : 0.1 µrd/mm (locally?)

Two parallelisms:Bragg planeCross plane

koHzu situation:Lower cam “removes” the weight(not the angular deviation)

BM23 PARALLELISM ERROR INVESTIGATIONS

15

20

25

30

35P

aral

lelis

m e

rror

(ar

csec

.)

initial setup with setup of the translation with setup of the translation and pre loading (without load) with setup of the translation and pre loading (with load) 27/09/2010 Kohzu reference

Before upgrade

Courtesy J. Borrel, O. Mathon

AutocollimatorMeasurements

Room temp.

5 10 15 20 25 30-5

0

5

10

15

Par

alle

lism

err

or (

arcs

ec.)

tetaB (deg)

As delivered several year before

Final performance

6000

8000

10000

12000

14000

16000

18000

pIE

ZO

PI c

ontr

oler

val

ue (

A.U

.)

data 2011 data 2003 before Kohzu intervention data 2007 after Kohzu intervention

In situPiezo correction

PiezoController

At present

Room temp.

KoHzutuning


Seems to be a consequence of the weight added on the Kohzu mechanism3 crystals set instead of 2

0 10 20 30 40

-8000

-6000

-4000

-2000

0

2000

4000

6000

pIE

ZO

PI c

ontr

oler

val

ue (

A.U

.)

angle (deg.)

Controllervalues

tuning

REMARK: RADIAL MOVE MAY BE OF LESS IMPORTANCE

Remark:Radial stroke ensure fixed exit

But

Fixed exit is not mandatory for energy scanning ∆E ~1000 eV

∆θΒ2nd Radial stroke

Sample slitOr

Pseudo-Fixed exit operation

Beam moves

Used when experimentRequires large size beam(rather seldom)


Or seg. slit

(rather seldom)

Slit distance

should remainsmall

Angular apertureShould ≤ 100 µrd

CRYSTAL PARALLELISM DURING ENERGY SCANS

Parallelism error ≤ 100 nrd during scans

The error is lower than the limit Strategy when the error is above the limit ( 100 nrd)

Metrology in situ possible. watch

Error is repeatable(RT + cryo) Error is not repeatable

ProbablyProbe system

Corrected within Mono♦ Fine stages♦ Probes (Look-up table+ Piezo strain gauges)

Corrected in real time♦ Probe system (straight edge+ capacitor)♦ Dynamic TF with piezos

No correction(ideal case!)

Corrected externally♦ probes but no feedback♦ at the sample (vert. + Horz. Positioning)


Recent monoMay be there

Minimum requirementsWith still a simple system

Complex system Would require developments

CRYSTAL PARALLELISM : AVOID MECHANISM THERMAL SHRIN KAGE

Exp: A. Chumakov, K. Martel (ESRF ID18 – ID20)

Solid ceramic isolator (very low conductivity)

High stiffness modulus (no vibrations)

Limited conductive contact areas (no thermal leaks)

Then crystal travel stages


Then crystal travel stages

Passive supportstructure

Corrective stages

ESRF wish

Exp at ESRF with EUROTHERM temperature PID controller At ID20 ± 1 m°K for post mono

Courtesy C. Henriquet

THE BRAGG AXIS MOTOR CASE

Motoroperation

Motorscans


LN2 FEEDER BOX: A “HEAVY INFLUENCE” , MAY BE AWAY F ROM SUPPLIER’S

External flexible line

Moving LN2 feeder boxSeems quite stronglyInfluencing theMass inertia


BRAGG AXIS DYNAMICS

More Tricky windingLN2 feeders Fixed box

Moving box

Bragg axis motor should move an “optimized” set

LN2 feeders Fixed boxMoving box

Here:40 ≤ I

kg.m2≤ 90


5%8%85%

Fraction of mass inertia sharing

LN2 feed boxWith flexible

windings

BRAGG AXIS DYNAMICS

Mass centroidFar from axis


Adding a counterweight improves Unpowered reaction but magnifiesThe dynamic considerations

Typically from 40 to 60 Kg.m2

Added external Counterweight:30 Kg @ 700 mm

MOTOR DYNAMIC OPERATION

J is the total mass inertia of the Bragg axis

Motor torque = Dynamic load + Permanent load• Friction• Off axis masses• Hand out forces• Everything at constant velocity

• Everything• Connected with• acceleration

idtd CJtC += 2

2

)( θJ is the total mass inertia of the Bragg axis

Ci is anything constant

C a t β, i, I,( ) J m i⋅ dθ2 t β,( )⋅1

i η⋅I dθ2 t β,( )⋅ C o+ C f+( )+:=

Everything permanentEverything dynamic

Co : off axis masses torque

Cf : Friction torque(Ferrofluidic + bearings)


“slow mvt”“fast mvt”

Friction &Permanentresistance

Motor acceleration is magnifiedBy the gearbox

Everything is divided by the gearbox(with efficiency losses η)

BRAGG AXIS : REQUIRED TORQUE FOR ENERGY SCANNING

Required torque for scanning [Nm]

TorqueIn Nm

Direct drive technologyHigh gearbox cascade

High motor powerdissipation

Gearbox reduction ratio: i (no unit)

In Nm cascader1 r2 run by steppers or brushless

Few casesthere !

Shifts that wayare due to acceleration change

i=1 means direct drive operation


Gearbox reduction ratio: i (no unit)

Solving resolution withFine motor steps

Solving resolution with highDensity motor steps

“Motor overcomes encoder”

i=1 means direct drive operation

“Encoder overcomes motor”

… attractive idea: Mechanically more simple

A typical gear box

SPACE – VELOCITY – ACCELERATION – AND TORQUE

Angular triangular scansIn degrees

2%

The acceleration of the scan is a fraction in % of the scan angle here 1°

5% 10% Acceleration of a full cycle at 5% ramping

Space plot

Angular Velocity in °/s

Time in sCorresponding dynamic torque of this 5% ramping

Nm

Time in s

Direct drive motor torque Motor 2


Time in s

Time in s

Motor torque with gearbox Motor 1

Where the basic relation

J is the total mass inertia of the Bragg axis

idtd CJtC += 2

2

)( θ

STABILITY ISSUES: OUTLINES

CrystalsTimeconstantCompton

ScatteringIsolatedmechanisms


FlowInducedVib.

CrystalCagedesign

Monodeforms

Architecturalanalysis

GLOBAL VIBRATION CURING

Crystal mount pbs


Vibration cure:

• “Pillar like jacks” removed everywhere• Base block bonded to floor

Typical stand weaknesses

Older BM29 mono – 2008 – Courtesy M. lesourd

LATERAL MOVE TRANSFER FUNCTION: UPGRADING A KOHZU ST AND AT ID14

DSP response

Lateral move transfer function HighValues !

Lateral displacement transfer function TFWith respect to floor sourceresponse DSPTFDSP 2= DSP

source

y2

LowValues !


Bonding to floor

TF=

y1

y2

y1 Adjusting at higher level

Travels as high as possible

TWO MONOS IN TEST AT THE ESRF ID6- (IN 2008)

Elec noise

TransferFunctionrecords

New DCM for spectroscopy an engineering challenge - 13 May 2014 - Y. Dabinl Title of Presentation l Date of Presentation l Au thorPage 25

Pillarshere

WedgeJackshere

wedge

FLEXURE DESIGN: VERY HARD TO BE PERFECT MODAL ANALY SIS

8 KgVirtual mass

Mode at 21 Hz Mode at 31 Hz Mode at 36 Hz

Flexures: Difficult to make non-coupled movements

Simulation of a 2nd crystal positioning


Flexures: Difficult to make non-coupled movements

Ideally: design single degree of freedom with piezo.Done for white beam optics (ESRF upgrade)

Vibration wise: design target should be at 130 Hz

Necking is deformingBeamlines: Not enthusiastic to flexures

SINGLE DOF FLEXURE BASED GONIOMETER

The problem to address is the availability of single DoF high resolution goniometer(low parasitic motion)

30 Kg rocking mass

Fine rocking :Rx:0

130Hz loaded at 30 KgWhite beam mirrorESRF upgradeNot easy to implement forA crystal set


Fine rocking :• Rx:0• Ry:0• Rz: 15 nrd• low parasitic motions

SCATTERED POWER FROM BM5 CRYSTAL

BM Spectrum150 W (per 1mrad horizontal divergence)

Scoring plane

Spatial distribution over the scoring plane

5 mm

New DCM for spectroscopy an engineering challenge - 13 May 2014 - Y. Dabin

50 mm

10 cm thick

RESULTS: SCATTERED POWER ON FIRST CRYSTAL

Bragg angle [deg] scattering [%]

0.17 23.58

1 12.57

2 10.3

3 8.67

5 7.51

10 6.2215

20

25

pow

er [%

]

High energy position

10 6.22

15 5.6

20 4.72

30 3.69

40 3.13

60 2.62

70 2.53

80 2.79

90 2.68 0

5

10

15

0 20 40 60 80 100

Bragg angle [deg]

Sca

ttere

dpo

wer

[%]


The scattered power (fraction of energy irradiated) has been calculatedusing the incident white bending magnet spectrum shown.It is always less than 25% and less than 5-8% for usual Bragg angles (> 3 deg)

The scattered power has been calculated using “Penelope” codeHere input power is 150 W

45° about50 mm projection

About 130mm


45° about

Example of a 30° Bragg incidenceOn a 1 mrd beam width at 30m150 W input beam

TYPICAL SCATTERING SHIELDING

Global sideLocal tungsten screens are 1 mm

Ceramic thermalbreak


Global sideScreenShielding thePositioningMechanisms5 mm copper

Local tungsten screens are 1 mm

Global sideScreen

Local screenFor Xtals


ScreenShielding thePositioningmechanisms

Global screen is 5mmLocal screens are 1 mm tungsten

NEW DCM FOR SPECTROSCOPY AN ENGINEERING CHALLENGE R EVIEW

Shielding

Heater Always a goodConfig.

We a far from being blockedStill many issues

Good dynamics analysis

Labs have best operation analysisWe are the users

Direct drive

Gearbox motor

Good stage

All fixed

Labs have best experience forStability issue


Good plateHigh design validation

Thank you for your attention…

new dcm for spectroscopy an engineering...new dcm for spectroscopy an engineering challenge -13 may...

Documents